The Next Rembrandt

In the news recently you may have heard talk of a ‘new’ Rembrandt. They weren’t referring to some young prodigy setting the art establishment ablaze, nor about a rediscovered classic from the master himself. They were talking about a truly remarkable achievement by the Delft University of Technology and the Rembrandt Museum. Their aim was to use data capture and algorithms to capture the essence of Rembrandt and then to create not a replica of an existing Rembrandt but a completely new and original one. You can see the result below.

rem

Now, I am a long way from being an art historian but I am a fan of Rembrandt and, to my untrained eye at least, this painting looks a heck of a lot like a Rembrandt. This shouldn’t come as a surprise considering how much effort went into producing this artwork. They have a beautiful website if you want to get the firsthand take on the project – not until you’ve read my musings, though, obviously.

The first step was to collect data. Luckily, Rembrandt was prolific and so there are plenty of his works out there from which to collect the necessary information. Looking at his main body of work they wanted to pick a subject for the painting. Although he did do some landscapes what he is really known for is his portraiture and so they decided to do a painting of a man. They scanned 346 paintings in fine detail capturing not just which colours were used where but also the texture and physical depth of the oil paints used.

They next used a series of algorithms to capture the exact architecture of a Rembrandt painting. What are the distances between the eyes, ears, nose and mouth. What shape are each of those features? The hairstyles and shape of the facial hair? What clothes were worn? What is in the background?

Having crunched the data they then gave their design software the parameters of its task. They asked it to create a portrait of a Caucasian male with facial hair, between the ages of thirty and forty, wearing black clothes with a white collar and a hat, facing to the right. That’s it. No other guidance was given. They set it on its way and after 500 hours of rendering the image above was the result.

Personally I’m not sure which is more impressive: the fact that they have so faithfully recreated Rembrandt’s style or that it looks like a real human. The proportions are perfect and don’t look remotely out of place; there is no uncanny valley effect here. You can easily imagine this guy wondering around a civil war reenactment or some such event. They weren’t finished with just a two dimensional digital image, however; one of the things I love about oil paints is the three dimensional depth that they have. Quite often when I’m in a gallery I will get as close as I can to the surface of a painting just to see the relief and the way that the paint was applied to the canvass.

The team incorporated three dimensional scan data into the 2D image so that it could be 3D printed onto the canvass using 13 layers of ink. This allowed them to replicate the brushwork to give the finished product the true ring of authenticity.

Now, what is not up for debate is whether or not this is an impressive technical feat. It is. If they had got it wrong they may have inadvertently produced a Picasso instead of a Rembrandt. But they didn’t; they got it spot on and did an outstanding job of capturing the warts-and-all, spotlit, soft focus style so characteristic of Rembrandt. Where there is certainly a debate to be had is in questions like: Is this art? Who owns the copyright? Does this add to the art world or detract from it?

Is it really possible to capture the human essence of portraiture using algorithms? Does it change your perception of the picture knowing that it was created ‘artificially’? I can certainly imagine some people railing against this type of art. If all the technical skill is taken out of being able to produce such an impressive piece does that devalue it? Or, on the other hand, does this democratise high art? Does this open art up to the masses who don’t have the skill to paint like this but might have wonderful ideas for composition that they’d otherwise never be able to realise?

Would you go to an exhibition compiled exclusively from artworks generated by computer? Perhaps Rembrandt doesn’t do it for you, but what if it was done in the style of Van Gogh, Titian, Caravaggio, Lichtenstein or whoever else your favourite might be? What if you could create something in their style to put on your living room wall?

There probably aren’t any right answers to these questions but I’ve certainly enjoyed considering them over the past week since I first saw this new Rembrandt and, for the record, I’d happily go see an exhibition of such artworks.

Advertisements

4 thoughts on “The Next Rembrandt

  1. Good questions; in my opinion, they should label it computer graphic art, crediting the team that put in the work. That would be until a truly AI machine is produced, whereupon the questions will arise again.

    Also, I think some questions may be answered by people’s reaction to the painting when it is displayed. I once attended a Van Gogh exhibit in SF at the De Young, where I saw a self-portrait he did. Standing at a certain distance from it, the face comes, in one’s perception, alive. I moved around to test it, and from that one point, the face literally looks right into your eyes and practically winks at you. It was uncanny, and I’ve only seen it in a couple other paintings…. Whether this one will be such a piece remains to be, shall we say, seen?

    Thanks, good one..

    Liked by 1 person

    1. I can go along with that. I’ve had some strong reactions to this one. At least one initially refused to accept this as art in any way until we agreed the credit can go to the coders. He argued that art has to be derived from human emotion. I don’t buy that. I think there’s art in nature that has nothing to with us. Or is that the distinction between art and mere beauty?

      Liked by 1 person

      1. Well, I suppose the concept of art can be flexible, according to whose concept it is. I’d agree that not all art is human, but, I would say only a human could recognize it, as we are the only ones we know of who do so. (though there are elephants and chimps who sell their paintings). Anyway, to believe we are the only ones capable of creating artistic beauty is absurd; just look at a spiderweb, each of which is an individual piece of beautiful artwork. To deny that is to deny one’s own perceptions…. I’d say the one who insists on that has some doubts as to their own adequacy, thus expressing jealousy, which no balanced person of intelligence would ever do.

        Beauty and art are two mutually connected concepts to me; not the same, but, art would not exist without beauty, I think, so it might be considered a more basic idea; art, while pretty all-encompassing, remains a search to create that beauty.

        Great subject for examination…. Gets into human motivation, naturally, I suppose….

        gigoid

        Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s